« first day (778 days earlier)      last day (1344 days later) » 

10:25 AM
3
Q: MathJax /newcommand in question titles is incorrectly scoped to the entire page, unlike in posts and comments

MartinThere are many sites in the network where MathJax is enabled. In MathJax we can also define macros - which can be used basically as shortcuts. However, some problems might be caused by such macros, since macros with the same name defined in different answers might clash.1 For this reason, since J...

@rene In the footnote at the bottom, there is a note about posts that depended on a buggy MathJax behavior that was fixed in January 2019. Is there a way to scan for those posts in SEDE? (cc @Martin)
 
 
4 hours later…
2:06 PM
I guess that a reasonable way would be to download the data dump and analyse the XML file. (I think it is better to do something like that by other tools, not in SQL.)
Nevertheless I have at least some queries which look for one specific macro - as you can see here: chat.stackexchange.com/transcript/88939/2020/9/13
We talked about this also in this room a bit: chat.meta.stackexchange.com/transcript/1223/2019/2/1
Feb 1 '19 at 19:56, by rene
@Martin well, no, not in one go at least. I'm not sure if MathJax would be suitable to be parsed with an regex but uptil SqlServer 2017 there is no support for REGEX in a where clause.
Feb 1 '19 at 19:57, by rene
In theory you could write a stored procedure that handles each row/field and then do the parsing in TSQL logic. That can be fun but is cumbersome as tsql by itself is not a language with many powerful language constructs.
@SonictheMaskedWerehog See above.
I hoped that I might find the time to write some program that goes through the xml in the SQL dump and analyses which macros are defined there and whether they are used also in other posts.
However, I did not get around to it - and in the following months I am probably going to be rather busy.
 
 
2 hours later…
4:08 PM
Not sure if I left this approach before: data.stackexchange.com/math/query/1293969 I doubt this is more efficient then doing a LIKE but if you need things like, is not followed by etc this might turn out to be quicker then writing a parser (cc @SonictheMaskedWerehog @Martin)
 

« first day (778 days earlier)      last day (1344 days later) »